It is important to have the capability to perform precise measurements of gene expression levels in tumor cells. Available technology permits the assessment of levels of gene expression in samples containing a minimum of 10,000 cells. Thus, there is a need for methods that will permit such measurements in samples containing fewer tumor cells. A specific aim of this proposal is to complement a new approach that will yield quantitative data on the relative concentration of specific mRNA molecules in tissue samples containing less than 200 cells. Sequence detection is accomplished on oligonucleotide microarrays, using a target-directed DNA ligation step coupled to a Rolling Circle Amplification (RCA) unimolecular detection system. Each target detection event generates a primer that can be amplified by RCA. Each amplified DNA molecule generated by RCA remains localized on the array surface, and is imaged as a discrete fluorescent signal, indicative of a specific molecular ligation event. Expression profiles are generated as histograms of single molecule counts. Additionally, the DNA ligation step can e adapted to the detection of mRNAs containing point mutations. This capability will be developed for detection of one somatic mutant mRNA molecule in a background of 1000 wild type mRNA molecules, using K- ras mutations as an experimental model. Arrays for the analysis of 100 different gene products will include mRNAs known to be up- or down- regulated in cancer cells, wild type or mutant K-ra mRNAs, and housekeeping genes. Adequate controls will be incorporated in the system to insure its reliability. Another specific aim is to evaluate the coupling of this highly sensitive detection technology with laser- assisted tissue microdissection. We will combine these two technologies to demonstrate the capability for high resolution tissue analysis in: (a) normal and cancerous prostate, and prostate tissue with varying degrees of prostatic intraepithelial neoplasia; (b) normal colonic epithelium, as well as adenomatous epithelium with varying degrees of dysplasia, and colonic adenocarcinoma. The capability for measuring gene expression levels in samples containing as few as 10-20 cells, together with the capability for detection of somatic point mutations at several loci known to be altered with high frequency, will provide the infrastructure to address the question of possible microheterogeneity in gene expression profiles in small clusters of cells in dysplasia and cancer.